Design, expression, purification, and application of novel recombinant miR-491 molecules to define the biogenesis and function of miR-491-3p versus -5p in posttranscriptional regulation of UDP-glucuronosyltransferase 1A1

Abstract

Interindividual variations in drug metabolism involve various factors, including posttranscriptional gene regulation mechanisms controlled by microRNAs (miRNAs or miRs) derived from the genome. The aim of this study was to use RNA bioengineering technology to produce novel recombinant human miR-491-5p, miR-491-3p, and pre-miR-491 molecules, namely BioRNA/miR-491-5p, BioRNA/miR-491-3p, and BioRNA/pre-miR-491, respectively, and define their functional difference in regulating UDP-glucuronosyltransferase 1A1 (UGT1A1) expression and drug-metabolizing capacity. All 6 BioRNAs were heterologously overexpressed in Escherichia coli (>30% of total RNA) and isolated by fast protein liquid chromatography to high purity (>97%). As BioRNA/pre-miR-491 agents were processed to both 5p and 3p strands in Hep3B and HepG2 cells, BioRNA/miR-491-5p and -3p were selectively processed to 5p and 3p, respectively, and each accumulated to greater levels. Immunoblotting and immunofluorescence studies demonstrated the efficacy of BioRNA/miR-491-3p to suppress UGT1A1 protein levels in Hep3B and HepG2 cells, localized on the endoplasmic reticulum, exhibiting monomeric (∼55 kDa) and oligomeric (∼150 kDa) bands under different conditions, whereas BioRNA/pre-miR-491 and miR-491-5p had no effects. Using a fluorescent substrate, N-butyl-4-(4-hydroxyphenyl)-1,8-naphthalimide, lower UGT1A1 drug-metabolizing capacities were found in cells treated with BioRNA/miR-491-3p. In addition, liquid chromatography-tandem mass spectrometry analysis revealed a 45% reduction of estradiol 3-glucuronidation activity by BioRNA/miR-491-3p in Hep3B cells, whereas formation of estradiol 17-glucuronidation mediated by other UGTs was unchanged. Together, these results underline the role of miR-491-3p in regulating UGT1A1 and its impact on cellular drug-metabolizing capacity while demonstrating the applications of recombinant miRNA agents to delineating the importance of posttranscriptional gene regulation in drug metabolism.

Significant Statement

Research on posttranscriptional gene regulation mainly uses miRNA mimics chemically synthesized in vitro. This study successfully produced 6 novel recombinant miR-491 molecules through in vivo fermentation with transfer RNA scaffold and transfer RNA-fused pre-miRNA carrier-based technologies, which were further utilized to delineate the biogenesis and function of miR-491-3p versus -5p in modulating UDP-glucuronosyltransferase 1A1 protein levels and drug-metabolizing capacity. The findings demonstrate the role of miR-491-3p in regulating UDP-glucuronosyltransferase 1A1 and value of recombinant miRNA agents for studying drug metabolism.

Graphical abstract

1. Introduction

UDP-glucuronosyltransferases (UGTs) are critical phase II drug-metabolizing enzymes that catalyze glucuronidation reactions facilitating the elimination of many endogenous compounds, such as bilirubin and steroid hormones, as well as various xenobiotics including acetaminophen, estradiol, tamoxifen, and irinotecan.1, 2, 3, 4 UGT enzymes are highly expressed in metabolic organs, especially the liver and intestine, and UGTs are located on the endoplasmic reticulum (ER) within cells. Human UGTs are comprised of 4 families, UGT1, UGT2, UGT3, and UGT8, including more than 20 members.² The most abundant UGT isoforms are often found in the UGT1A and UGT2B subfamilies, among them the UGT1A enzymes are known to metabolize a large number of therapeutic drugs.^2,5 The human UGT1A gene is located on chromosome 2q37, and alternative splicing leads to 9 UGT1A isoforms, each consisting of a unique first exon joined to the common exons 2-5 and 3′-untranslated region (3′UTR).6, 7, 8

UGT1A1 has been revealed as the most abundant UGT1A isozyme in human livers while exhibiting large interindividual variations9, 10, 11 that can significantly affect drug or endobiotic metabolism and subsequently, influence drug responses or cause diseases.^2,7,12 One important factor causing large interindividual variations is genetic polymorphism.^2,13, 14, 15 For example, the UGT1A1∗28 allele, prevalent in 30%–45% of Caucasians, 10%–20% of East Asians, and 42%–51% of Africans, is known to cause a lower UGT1A1 protein level and glucuronidation activity. Individuals carrying homozygous UGT1A1∗28 allele may have an impaired bilirubin clearance and increase the risk of conditions such as hyperbilirubinemia, Gilbert syndrome, and Crigler-Najjar syndrome.² Furthermore, people with the UGT1A1∗28 allele are more susceptible to irinotecan toxicity due to a lower glucuronidation activity and less efficient clearance of its active metabolite SN-38.^13,14 Although genetic polymorphism accounts for some variability, large differences in UGT1A1 mRNA and protein expression levels likely involve many other molecular mechanisms, such as variable gene regulatory factors and signaling pathways.^2,16, 17, 18

Recent studies have revealed that some epigenetic factors, such as noncoding microRNAs (miRs or miRNAs), may contribute to the regulation of UGT1A1 gene expression.19, 20, 21, 22, 23, 24 These genome-derived miRNAs are functional small RNAs governing posttranscriptional gene regulation (PTGR) through binding to complementary sequences within the 3′UTR of target mRNAs, resulting in translational repression or mRNA degradation.²⁵ Indeed, specific miRNAs, such as miR-491-3p,¹⁹ miR-141-3p,^20,23 and miR-214-5p,²⁴ have been identified to control the PTGR of UGT1A1 and some other UGT1A isoforms. Besides UGT1A1, many other drug-metabolizing enzymes, transporters, or their regulatory factors have been shown to be regulated by functional miRNAs that could further influence drug metabolism and disposition and chemosensitivity.^26,27 Research on the roles of miRNAs in PTGR of enzymes and transporters should advance the understanding of complex mechanisms behind interindividual variations in pharmacotherapy and provide insights into the development of new therapies.

Previous studies on miRNA-controlled PTGR mainly use chemically synthesized miRNA mimics with various chemical modifications which have proved their values. Nevertheless, miRNA mimics may not recapitulate the characteristics of natural miRNAs produced in cells.^28,29 Although the guide strand consists of the same primary sequence, miRNA mimics agents from different vendors undoubtedly contain different types and degrees of chemical modifications, and they are literally distinct molecules.^30,31 It is also obscure whether the double-stranded miRNA mimics with extensive chemical modifications are mainly processed to the guide strands of interest, either 5p or 3p, in comparison to natural miRNA duplexes that rarely have any posttranscriptional modifications. Moreover, it is unknown whether the passenger strand comprised of extensive chemical modifications would inhibit a guide strand of interest.³² Very recently, our lab has developed a novel RNA bioengineering technology that enables high-yield and large-scale, in vivo production of recombinant miRNA (BioRNA/miRNA) agents.33, 34, 35, 36 Starting from a transfer RNA (tRNA) scaffold to assemble pre-miRNA agents,^37,38 this technology employs specific tRNA-fused pre-miRNA carriers (eg, the hsa-pre-miR-34a) to accommodate target miRNAs^33,35,36 (Fig. 1A) which not only enables heterologous overexpression in Escherichia coli but also achieves the release of 5p miRNA predominantly in human cells (likewise, a 3p miRNA of interest can be adapted at the 5p position within the hsa-pre-miR-34a carrier; Fig. 1A). Our studies have demonstrated the utilities of novel BioRNA/miRNA molecules,^33,36,39, 40, 41 being an addition to current tools for miRNA research.

Fig. 1.

Design and production of new human miR-491 agents. (A) Schematic illustration of 6 different BioRNA molecules, including hsa-pre-miR-491 (light purple) assembled with human glycyl (purple) and leucyl (yellow) tRNA (htRNA^Gly and htRNA^Leu, respectively) scaffold, as well as miR-491-5p (red) and miR-491-3p (green) with htRNA-fused hsa-pre-miR-34a (blue) carrier. (B) Urea-PAGE analyses of total RNAs from bacteria transformed with BioRNA-expressing plasmids confirmed the heterologous overexpression of target BioRNAs. Total RNA from untransformed E. coli HST08 was used as control. (C) FPLC-UV trace during the purification of the representative BioRNA^Gly/miR-491-3p molecule. The insert shows urea-PAGE analysis of target RNA fractions collected during FPLC separation. (D) Urea-PAGE analyses of the purified BioRNA/miR-491 molecules. (E) HPLC determination of the purity of final BioRNA product (BioRNA^Gly/miR-491-3p is presented as an example, showing 99.0% homogeneity).

In this study, we aimed to design and produce a total of 6 recombinant pre-miR-491, miR-491-3p, and miR-491-5p molecules by using 2 different human tRNA scaffolds and their coupled hsa-pre-miR-34a carriers (Fig. 1A). After heterologous overexpression and purification, these BioRNA/miR-491 molecules were used to reveal the generation of miR-491-5p and -3p from pre-miR-491 in human liver cells, as well as selective release of miR-491-5p and -3p from BioRNA/miR-491-5p and -3p molecules, respectively. In addition, the functional differences between miR-491-3p and -5p to modulate UGT1A1 protein levels and drug-metabolizing capacity in human cells were demonstrated. These results establish the impact of miR-491-3p on UGT1A1 expression and glucuronidation metabolism, underscoring the value of recombinant miRNA agents.

2. Materials and methods

2.1. Chemicals and materials

DMEM medium (Cat# 11965092), EMEM medium (Cat# 10009CV), BEGM medium (Cat# CC3171), 0.05% trypsin-EDTA (Cat# 25052CI), PBS (Cat# 10010023), FBS, Opti-MEM (Cat# 31985070), bicinchoninic acid (BCA) Protein Assay Kit (Cat# 23227), Lipofectamine 3000 (Cat# L3000075), polyclonal anti-UGT1A1 antibody (Cat# PA5-76862), and estradiol (Cat# E8875) were purchased from Thermo Fisher Scientific. Direct-zol RNA MiniPrep Kit (Cat# R2072) was bought from Zymo Research, and primers were synthesized by Integrated DNA Technologies. TGX Stain-Free FastCast Acrylamide Kit, 10% (Cat# 1610183), Clarity Western ECL Substrates (Cat# 1705061), polyvinylidene difluoride membranes (Cat# 1620184), and iTaq Universal SYBR Green Supermix (Cat# 1725124) were purchased from Bio-Rad. Monoclonal anti-UGT1A1 antibody (Cat# A22609) was obtained from ABclonal Science, Inc. Anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (Cat# 2118L) was purchased from Cell Signaling Technology, Inc. Anti-β-actin antibody (Cat# A5441), β-estradiol 17-(β-D-glucuronide) sodium salt (Cat# E1127), 17-ethynylestradiol (Cat# E4876), and alamethicin (Cat# E1127) were obtained from MilliporeSigma. β-Estradiol 3-(β-D-glucuronide) sodium salt (Cat# HY-W585842) was purchased from MedChemExpress LLC. Uridine 5'-diphosphoglucuronic acid (UDPGA) (sodium salt hydrate) (Cat# 20674) was purchased from Cayman Chemical Company, Inc. N-butyl-4-(4-hydroxyphenyl)-1,8-naphthalimide (NHPN)42, 43, 44 was a gift from Drs Ling Yang and Guan-Bo Ge. The UGT Activity Assay/Ligand Screening Kit (Fluorometric) was purchased from Abcam (Cat# ab273331). All remaining chemicals, organic solvents, and materials were purchased from Thermo Fisher Scientific, VWR, or Sigma-Aldrich.

2.2. Design, cloning, expression, and purification of 6 different BioRNA/miR-491 molecules

Human pre-miR-491 was directly fused to human glycyl and leucyl tRNA by replacing anticodon sequences to offer BioRNA^Gly/pre-miR-491 and BioRNA^Leu/pre-miR-491, respectively (Fig. 1A), as described.^37,38 By contrast, the tRNA-fused hsa-pre-miR-34a carriers^33,35,36 were used to offer BioRNA^Gly and BioRNA^Leu versions of miR-491-5p and -3p molecules by replacing miR-34a duplex with payload miR-491-5p and -3p, respectively, along with respective passenger strands (Fig. 1A). The sequences of individual BioRNAs were provided in Supplemental Table 1, and BioRNA-expressing plasmids were constructed by inserting BioRNA coding sequences, amplified with specific primers (Supplemental Table 2), into the pBSTNAV vector linearized by restriction enzymes EcoRⅠ and PstⅠv, as described.^33,35

After confirming the insert through DNA sequencing, BioRNA heterologous expression and purification was conducted as reported.^33,35 Briefly, BioRNA-expressing plasmid was transformed into HST08 E. coli competent cells, and overexpression of target BioRNA was verified through urea-polyacrylamide gel electrophoresis (PAGE) analyses of total RNAs isolated from E. coli. BioRNAs were purified by an anion exchange fast protein liquid chromatography (FPLC) method on an NGC Quest 10 Plus Chromatography system (Bio-Rad). The purity of isolated BioRNA was semiquantitatively assessed by urea-PAGE analysis and quantitatively determined by high-performance liquid chromatography (HPLC) method on a Shimadzu LC-20AD HPLC system (Shimadzu Corporation). Endotoxin levels were determined by using a Pyrpgent-5000 Limulus Amebocyte Lysate kinetic assay kit.

2.3. Cell culture and transfection

Human hepatocarcinoma Hep3B, and HepG2 cell lines, human liver epithelial THLE-2 and THLE-3 cell lines, and mouse hepatocyte-derived AML12, and Hepa1-6 cell lines were purchased from American Type Culture Collection. Human Huh7 cell line was purchased from Japanese Collection of Research Bioresources. Mouse Hepa 54.3 cell line was purchased from Cytion. Hepa 54.3, AML12, Hepa1-6, and Huh7 were maintained in the DMEM medium, Hep3B and HepG2 were maintained in the EMEM medium, and THLE-2 and THLE-3 were maintained in the BEGM medium at 37 °C in a humidified atmosphere of 5% CO₂ and 95% air. All cell culture media were supplemented with 10% FBS (Thermo Fisher Scientific). Cells were seeded and incubated in cell culture plates overnight and then transfected with RNA with Lipofectamine 3000 reagent in reduced serum medium Opti-MEM, according to the manufacturer's protocols.

2.4. RNA isolation, reverse transcription, and quantitative real-time polymerase chain reaction analysis

Hep3B (500,000 cells/well) or HepG2 (750,000 cells/well) cells were seeded in 6-well plates, incubated overnight, and transfected with 20 nM of BioRNA/miR-491, control RNA, or vehicle for 72 hours with 3 biological replicates (N = 3/group) processed separately. Total RNA was extracted using the Direct-zol RNA isolation kit, and total RNA (300 ng) was used for cDNA synthesis with random hexamers or miR-491-5p and miR-491-3p specific stem-loop reverse transcription (RT) primers (Supplemental Table 3) using NxGen M-MuLV reverse transcriptase (Lucigen). Quantitative polymerase chain reaction (qPCR) analyses were performed on a CFX96 Touch real-time PCR system (Bio-Rad) using gene-specific primers (Supplemental Table 3) and iTaq universal SYBR Green supermix. Levels of mature miR-491-5p and -3p were normalized to U6 small RNA in corresponding samples by using the formula 2^-ΔΔCT, as described.^35,40

2.5. Protein isolation and Western blot analysis

Hep3B (500,000 cells/well) and HepG2 (750,000 cells/well) cells were seeded into 6-well plates and treated with 20 nM of BioRNA/miR-491 or control RNA agents. After 72 hours, cells were harvested and lysed in radioimmunoprecipitation assay buffer supplemented with protease inhibitors. Protein concentrations were determined using a BCA Protein Assay Kit. Equal amounts of proteins (20–30 μg per well), either boiled (95 °C for 9 minutes) or unboiled, were separated on 10% SDS-PAGE gels and transferred onto polyvinylidene difluoride membranes. Membranes were blocked in 5% nonfat milk for 2 hours at room temperature and then incubated overnight at 4 °C with polyclonal primary antibodies against UGT1A1 (1:700 dilution), monoclonal primary antibodies against UGT1A1 (1:4,000), β-actin (1:3,000), or GAPDH (1:1,000), respectively. After washing, membranes were incubated with horseradish peroxidase-conjugated secondary antibodies (anti-mouse, 1:3,000, Cell Signaling Technology; anti-rabbit, 1:10,000, Jackson ImmunoResearch Europe Ltd) for 2 hours at room temperature. Protein bands were visualized using enhanced chemiluminescence and imaged with a ChemiDoc MP system (Bio-Rad). Band intensities were quantified using Image Lab software and normalized to total protein or respective β-actin or GAPDH levels. Experiments were performed in biological triplicates (N = 3 per group).

2.6. Immunofluorescence confocal imaging study

Hep3B or HepG2 cells were seeded in an 8-well chamber slide (50,000 cells/well) and treated with 20 nM of recombinant miR-491-3p or control RNA for 72 hours. To determine the localization and expression level of UGT1A1, live cells were first incubated with ER-tracker (1.5 μM) in PBS for 1 hour and then fixed with 4% paraformaldehyde. After blocking with 3% BSA in PBS, cells were incubated overnight at 4 °C with a primary polyclonal UGT1A1 antibody (1:300). This was followed by a 1-hour incubation with an anti-rabbit Alexa Fluor 488-conjugated secondary antibody (1:1,000; Cell Signaling Technology Cat# 4414S). In addition, a separate batch of cells was stained for F-actin using Alexa Fluor 680 Phalloidin (1×; Thermo Fisher Scientific Cat# A22286) according to the manufacturer’s protocol. Mounting reagent with DAPI (17 μL/well; VECTASHIELD, Cat# H-1200) was added to cells to stain the nuclei before imaging. Images were obtained by using a Leica Stellaris 5 Confocal Microscope platform with 10× and 63× objectives and Leica Application Suite software (Leica Microsystems). Experiments were performed in biological triplicates (N = 3 per group), and representative data are presented.

2.7. UGT1A1-mediated NHPN metabolism

Hep3B or HepG2 cells were seeded in 24-well plates at a density of 150,000 cells/well and treated with 20 nM of BioRNA/miR-491-3p or control RNA. After 72-hour treatment, the medium was discarded and cells were washed twice with PBS, then incubated with 10 μM of NHPN at 37 °C for 30 minutes to assess UGT1A1-mediated glucuronidation activity. In particular, cells were washed 3 times with PBS, and the fluorescence intensity of the metabolite NHPN glucuronide was immediately measured at excitation 370 nm and emission 520 nm^42,43 by using a SpectraMax iD5 microplate reader. Following the microplate measurement, live-cell imaging was performed using a Leica Stellaris 5 confocal microscope (10× objective with 2.5× zoom factor) to visualize intracellular fluorescence distribution. Finally, to ensure data accuracy, all fluorescence intensity values were normalized to protein concentrations in corresponding samples as determined by a BCA assay.

2.8. Estradiol glucuronidation and liquid chromatography tandem mass spectrometry analyses

Hep3B cells (3,000,000 cells/dish) were seeded in 100-mm dishes and transfected with 20 nM of BioRNA^Gly/miR-491-3p or control RNA for 72 hours in 4 biological replicates (N = 4 per group) that were processed in parallel. Cells were harvested after washes with ice-cold PBS for 3 times and then transferred to 1.5-mL tubes. Cell pellets were collected after centrifugation at 3000 g for 5 minutes, and 450 μL of 50 mM Tris-HCl buffer with 150 mM KCl (pH 7.4) was used for cell homogenization. The protein concentrations were then measured using the BCA assay.

The incubation was conducted in 100 mM Tris-HCl solution with 5 mM MgCl₂, consisting of cell homogenate (0.4 mg protein/mL), alamethicin (10 μg/mL), estradiol substrate (100 μM), and UDPGA (5 mM), similar as those described.45, 46, 47 200 μL cell of homogenate was mixed with 50 μL of alamethicin and placed on ice for 15 minutes for pore formation before adding 50 μL of estradiol. Pooled human liver microsomes (Coded H161, mixed gender; BD Discovery Labware, Inc.) were incubated under the same condition to serve as a positive control. The reaction was initiated by the addition of 100 μL of UDPGA after being warmed at 37 °C for 5 minutes. After 1 hour, the reaction was stopped by the addition of an equal volume of ice-cold acetonitrile. For the analysis of glucuronides, 400 μL of sample was mixed with 450 μL of acetonitrile containing 50 μM of 7-hydroxycoumarin (internal standard 1, IS₁), vortexed, and then centrifuged at 13,000g for 10 minutes. The resulting 800 μL supernatant was dried under vacuum and reconstituted with 100 μL of the initial mobile phase, and 10 μL was injected for liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. To analyze estradiol, 400 μL of sample was added with 400 μL of acetonitrile containing 50 μM of ethinylestradiol (internal standard 2, IS₂). The mixture was vortexed and then centrifuged at 13,000 g for 10 minutes, and 10 μL of the supernatant was directly injected for LC-MS/MS analysis.

Estradiol and its 3- and 17-glucuronide metabolites were quantified using an LC-MS/MS method adapted from those reported^46,48,49 on a 4000 QTRAP tandem mass spectrometry (AB Sciex) coupled with a Prominence Ultra-Fast Liquid Chromatography system (Shimadzu Corporation). Analytes were separated on a Waters Xbridge C18 column (3.5 μm, 2.1 × 50 mm) at a flow rate of 0.6 mL/min. To quantitate estradiol-3-glucuronide and estradiol-17-glucuronide, samples were eluted with mobile phase A (10 mM ammonium acetate with 0.1% formic acid) and mobile phase B (acetonitrile): 20% B (0–0.5 minutes), 20%–95% B (0.5–2.5 minutes), 95% B (2.5–3.5 minutes), 95%–20% B (3.5–3.7 minutes), and 20% B (3.7–5.2 minutes). To analyze estradiol, samples were separated with the mobile phase A (10 mM ammonium acetate) and mobile phase B (acetonitrile). The gradient elution program was as follows: 20% B (0–0.5 minutes), 20%–95% B (0.5–2 minutes), 95% B (2–4 minutes), 95%–20% B (4–4.2 minutes), 20% B (4.2–5.5 minutes). All analytes were detected in negative electrospray ionization mode using multiple reaction monitoring, m/z 447.1 → 271.0 for estradiol-3-glucuronide, m/z 447.1 → 271.0 for estradiol-17-glucuronide, m/z 161.1 → 133.1 for IS_1, m/z 271.0 → 144.9 for estradiol, and m/z 295.0 → 142.8 for IS₂. The linear calibration ranges were from 0.5 to 100 nM for both estradiol-3-glucuronide and estradiol-17-glucuronide, and 1–200 μM for estradiol. Accuracies were within 10% and precisions were within 15%. The recovery of the analytes was consistent and reproducible (variation coefficients, ≤ 5%), and the matrix effect was less than 10%. All data were acquired and analyzed using Analyst 1.6.3 (AB Sciex).

2.9. Statistical analysis

All values are mean ± SD. Student's t tests or one-way ANOVA with Bonferroni post hoc tests were used to assess the differences (Prism 8.0, GraphPad Software). Statistical significance is noted when P < .05.

3. Results

3.1. Design and production of new recombinant pre-miR-491, miR-491-5p, and -3p agents

To evaluate the biogenesis of miR-491-5p versus -3p strands, we designed 2 recombinant pre-miR-491 molecules with human tRNA^Gly and tRNA^Leu scaffolds^35,37 to offer BioRNA^Gly/pre-miR-491 and BioRNA^Leu/pre-miR-491, respectively (Fig. 1A). To delineate the functional difference between miR-491-5p and -3p, we employed respective tRNA-fused hsa-pre-miR-34a carriers^34,35 to assemble BioRNA^Gly and BioRNA^Leu versions of miR-491-5p and -3p molecules (Fig. 1A). After successful cloning of 6 target BioRNA-expressing plasmids, individual plasmids were transformed into the HST08 E. coli strain to achieve heterologous overexpression, as manifested by the appearance of new strong RNA bands around expected sizes (150–200 nt; Supplemental Table 1) following urea-PAGE analyses (Fig. 1B).

To isolate target BioRNA, total bacterial RNA was subjected to anion exchange FPLC purification. Fractions of the BioRNA as monitored by UV detection were collected and further verified by urea-PAGE analysis (Fig. 1C). The purity of final desalted BioRNA product was determined by 2 methods, semiquantitatively by urea-PAGE analysis that showing as a single RNA band (Fig. 1D) and quantitatively by the HPLC analysis that showing a single strong peak (Fig. 1E). All 6 BioRNA/miR-491 molecules purified by FPLC exhibited a high homogeneity (>97%; by HPLC) and high yield (>2.5 mg BioRNA from 250 mL bacterial culture) (Table 1). In addition, the endotoxin levels of each final BioRNA product were less than 5 Endotoxin Unit (EU)/μg RNA) (Table 1). Overall, all 6 recombinant BioRNA/miR-491 molecules were successfully produced in milligram quantity and high purity.

Table 1.

Yields and purities of new recombinant miR-491 molecules produced by using the BioRNA technology

BioRNA	Yield (mg BioRNA/250 mL culture)	Purity (%; by HPLC)	Endotoxin Level (EU/μg RNA)
BioRNAGly/pre-miR-491	4.53	98.5	4.88
BioRNAGly/miR-491-3p	7.00	99.0	3.20
BioRNAGly/miR-491-5p	7.20	99.0	4.26
BioRNALeu/pre-miR-491	2.79	97.9	1.46
BioRNALeu/miR-491-3p	12.5	97.1	4.10
BioRNALeu/miR-491-5p	9.62	97.7	1.62

3.2. Comparison of endogenous UGT1A1 protein levels in multiple human and mouse liver cell lines

To identify suitable cell lines for studying the biogenesis and functional difference between miR-491-3p and -5p in regulating UGT1A1, we compared the basal protein levels of UGT1A1 in multiple cell line models, including human hepatocarcinoma Hep3B, HepG2, and Huh7 cells and human epithelial THLE-2 and THLE-3 cells, as well as mouse liver-derived Hepa54.3, AML12, and Hepa1-6 cells (Supplemental Fig. 1). Although UGT1A1 proteins were readily detected in human Hep3B, HepG2, and Huh7 cells by western blots with 2 different antibodies, major bands around 150 kDa were surprisingly found in unboiled samples, and bands around 55 kDa (close to theoretical molecular weight of human UGT1A1) appeared after samples were boiled. Overall, the UGT1A1 protein bands were stronger in Hep3B and HepG2 cells. Moreover, the polyclonal anti-UGT1A1 antibody from Invitrogen (cat. #PA5-76862) seemed relatively more selective than the monoclonal anti-UGT1A1 antibody from ABclonal (cat. #A22609) as the latter led to multiple positive bands in all the mouse cell samples (Supplemental Fig. 1). Therefore, we selected human Hep3B and HepG2 cells as well as the polyclonal anti-UGT1A1 antibody for the following studies.

3.3. Release of miR-491-5p and -3p from different BioRNA/miR-491 molecules in human Hep3B and HepG2 cells

To examine the biogenesis of 5p and 3p strands from recombinant miR-491 agents, selective stem-loop RT-qPCR assays were employed to quantify the mature miR-491-5p and -3p levels in Hep3B and HepG2 cells following various treatments. The results demonstrated that, although BioRNA/pre-miR-491 was processed to the 5p strand mainly and 3p to some degrees in both Hep3B and HepG2 cells, leading to approximately 2- to 10-fold greater increase in the 5p strand than 3p, the BioRNA/miR-491-5p and -3p molecules were selectively processed to miR-491-5p and 3p, respectively (Fig. 2). Further, the miR-491-5p levels were about 3- to 5-fold higher in cells transfected with recombinant miR-491-5p than pre-miR-491, and miR-491-3p levels were 5- to 50-fold higher in cells treated with recombinant miR-491-3p than pre-miR-491, indicating the advantage of hsa-pre-miR-34a carriers. In addition, the BioRNA^Gly and BioRNA^Leu version of miR-491 molecules showed similar results in the release of mature miR-491-5p or -3p in both cell lines, suggesting that tRNA scaffold does not alter the miRNA biogenesis from BioRNA agents.

Fig. 2.

Release of the miR-491-5p and -3p strands from individual BioRNA molecules in Hep3B (A) and HepG2 (B) cells. Cells were treated with 20 nM of RNA or vehicle for 72 hours. The levels of miR-491-5p and -3p were determined by selective stem-loop RT qPCR methods and normalized to U6 levels in respective samples. Vehicle control group was set as 1.0. Values are mean ± SD. ∗∗∗P < .001; ∗∗P < .01; ∗P < .05; and ns, not statistically significant; as compared with corresponding control RNA treatments (one-way ANOVA with Bonferroni post hoc tests).

3.4. BioRNA/miR-491-3p molecules effectively modulate UGT1A1 protein levels in Hep3B and HepG2 cells

As human UGT1A1 has been revealed to be regulated by miR-491-3p albeit protein levels were not examined,¹⁹ we investigated the functional difference of BioRNA/miR-491-3p, pre-miR-491, and miR-491-5p in modulating UGT1A1 protein outcomes in cells. Our data showed that, compared with control RNA treatments, both BioRNA^Gly/miR-491-3p and BioRNA^Leu/miR-491-3p effectively reduced the UGT1A1 protein levels in both Hep3B (Fig. 3) and HepG2 (Fig. 4) cells, either the bands around 150 kDa in unboiled samples or single band around 55 kDa in boiled samples were examined. By contrast, pre-miR-491 and miR-491-5p did not alter UGT1A1 protein levels in cells. As the degrees of reduction in UGT1A1 protein levels by miR-491-3p showed some variations (Figs 3 and 4), slightly greater effects were found in Hep3B cells (eg, 20%–50% reduction by BioRNA^Gly/miR-491-3p) than HepG2 cells (20%–30%). In addition, the BioRNA^Gly/miR-491-3p demonstrated a little more consistent reduction in UGT1A1 protein expression across unboiled and boiled conditions, when compared with BioRNA^Leu/miR-491-3p.

Fig. 3.

Regulation of UGT1A1 protein outcomes by recombinant miR-491-3p molecules in Hep3B cells. Cells were transfected with 20 nM of BioRNA^Gly (A and B) or BioRNA^Leu (C and D) versions of miR-491 agents or control RNA for 72 hours, and both unboiled (A and C) and boiled (B and D) cell lysates were subjected to western blot analyses of UGT1A1 protein expression levels. UGT1A1 protein band intensities were normalized to corresponding β-actin, GAPDH or total protein levels, and control RNA groups were set as 1.0. Values are mean ± SD (N = 3 biological replicates per treatment group). ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001 as compared with control RNA (one-way ANOVA with Bonferroni post hoc tests).

Fig. 4.

Recombinant miR-491-3p agents modulate UGT1A1 protein levels in HepG2 cells. Cells were transfected with 20 nM of BioRNA^Gly (A and B) or BioRNA^Leu (C and D) versions of miR-491 or control RNA for 72 hours, and both unboiled (A and C) and boiled (B and D) cell lysates were subjected to western blot analyses of UGT1A1 protein expression levels. UGT1A1 protein band intensities were normalized to corresponding β-actin, GAPDH or total protein levels, and control RNA groups were set as 1.0. Values are mean ± SD (N = 3 biological replicates per treatment group). P < .05, and ∗∗P < .01 as compared with control RNA (one-way ANOVA with Bonferroni post hoc tests).

Moreover, we conducted immunofluorescence (IFC) confocal imaging studies to define the changes in subcellular UGT1A1 expression upon the treatment with recombinant miR-491-3p molecules. Consistent with the immunoblotting results, IFC studies showed that both BioRNA^Gly and BioRNA^Leu versions of miR-491-3p agents reduced UGT1A1 protein levels in both Hep3B and HepG2 cells, as manifested by obviously lower UGT1A1 fluorescence in corresponding cells treated with control RNA (Fig. 5; Supplemental Fig. 2). To verify UGT1A1 subcellular localization, cells were stained with ER-Tracker which showed colocalization with UGT1A1 signals. Together, these results demonstrate the effectiveness of recombinant miR-491-3p molecules to regulate the protein levels of endogenous UGT1A1 in Hep3B and HepG2 cells that are located on ER.

Fig. 5.

BioRNA/miR-491-3p modulates UGT1A1 protein levels in human Hep3B (A) and HepG2 (B) cells. IFC studies confirmed UGT1A1 (green) localization on the ER (red), whose intensities were obviously reduced by recombinant miR-491-3p molecules (20 nM for 72 hours) in both Hep3B (A) and HepG2 (B) cells. Scale bar, 100 μm.

3.5. Impact of BioRNA/miR-491-3p on UGT1A1-catalyzed NHPN metabolism in Hep3B and HepG2 cells

After observing an effective suppression of UGT1A1 protein levels by BioRNA/miR-491-3p, we further utilized a probe substrate, NHPN, to determine the impact of miR-491-3p on UGT1A1-mediated xenobiotic metabolism in cells. Glucuronidation of NHPN triggers a fluorescence “off-on” response, and the fluorescent intensity is proportional to the amount of NHPN-glucuronide produced by UGT1A1, enabling a sensitive detection and quantification of UGT1A1 enzymatic activity.^42,44 Our results showed that, compared with control RNA treatment, fluorescence intensities decreased 20%–30% in following the treatment with BioRNA^Gly/miR-491-3p (Fig. 6). On the other hand, BioRNA^Leu/miR-491-3p marginally reduced the fluorescence intensities in both cell lines, and it is not statistically significant (Fig. 6). These results illustrate the influence of BioRNA^Gly/miR-491-3p on UGT1A1-mediated NHPN glucuronidation in Hep3B and HepG2 cells, associated with the decrease of UGT1A1 protein levels (Fig. 3, Fig. 4, Fig. 5).

Fig. 6.

Impact of miR-491-3p on UGT1A1-mediated xenobiotic metabolism in human liver cells, as examined with the probe NHPN. Hep3B (A) and HepG2 (B) were treated with 20 nM of RNA for 72 hours and then incubated with 10 μM NHPN for 30 minutes. After live cell images were acquired, the fluorescent intensities of NHPN glucuronides were measured with a plate reader (excitation at 370 nm and emission at 520 nm) and then normalized to total protein concentrations. Values are mean ± SD (N = 3 biological replicates per group). ∗∗P < .01; and ns, not significant; as compared with control RNA (Student’s t-test).

3.6. Recombinant miR-491-3p reduces the UGT1A1 specific, estradiol 3-glucuronidation capacity in Hep3B cells

Given a consistent and relatively greater impact of BioRNA^Gly/miR-491-3p on UGT1A1 protein outcomes and NHPN glucuronidation (Fig. 3, Fig. 4, Fig. 5, Fig. 6), we thus investigated to what degree estradiol glucuronidation, a widely used UGT1A1 probe substrate,^45,46,48 would be altered by BioRNA^Gly/miR-491-3p in Hep3B cells. Estradiol can be metabolized selectively by UGT1A1 to estradiol-3-glucuronide, whereas estradiol-17-glucuronide is generated by other UGT enzymes, including UGT1A3 and UGT2B7 (Fig. 7A). As measured by accurate LC-MS/MS methods, the formation of estradiol-3-glucuronide was reduced about 45% by BioRNA^Gly/miR-491-3p in Hep3B cells as compared with control RNA (Fig. 7B). Meanwhile, the production of estradiol-17-glucuronide as well as the levels of estradiol remaining in the incubation mixtures showed no difference between miR-491-3p and control RNA treatments (Fig. 7B). These results demonstrate the role for miR-491-3p in the modulation of cellular UGT1A1-controlled drug-metabolizing capacity.

Fig. 7.

Recombinant miR-491-3p alters UGT1A1-mediated estradiol 3-glucuronidation capacity in Hep3B cells. (A) Schematic illustration of estradiol 3- and 17-glucuronidation catalyzed by different UGT isoforms. (B) Levels of glucuronides produced from estradiol in cell lysates after transfected with miR-491-3p and control RNA for 72 hours, as quantified by LC-MS/MS method. Values are mean ± SD (N = 4 biological replicates per group). ∗∗∗P < .001; and ns, not statistically significant as compared with the control RNA group (Student’s t-test).

4. Discussion

Research on miRNA-governed PTGR mechanisms and the development of RNA therapeutics primarily rely on RNA analogs chemically synthesized in vitro that differ significantly from RNA molecules made in vivo.^28,29 Although chemical modifications may improve RNA metabolic stability and PTGR efficacy, synthetic miRNA mimics from different vendors are literally comprised of distinct degrees and types of chemical modifications at various locations that are proprietary or confidential information undisclosed to the investigators, whose effects on miRNA biogenesis and functions are likely variable. In addition, it is obscure if the guide or dominant strand within the miRNA mimics would act as an antagonist or inhibitor of the passenger or minor strand, which may complicate studies on the minor strand of interest. Being an addition to current tools, recombinant miRNA agents produced in vivo by using newly developed BioRNA technology33, 34, 35, 36 have proved their values in miRNA research and development. In this study, we successfully applied this strategy to design, express, and purify 6 novel recombinant miR-491 molecules, supporting the robustness of BioRNA platform technology.

Utilizing these highly purified molecules, we first demonstrated the release of both miR-491-5p and -3p from pre-miR-491 in human cells where the miR-491-5p was more prevalent, consistent with current knowledge (https://www.mirbase.org/hairpin/MI0003126),⁵⁰ which are probably due to the processing of pre-miR-491 and stabilities of the 5p and 3p strands. On the other hand, the selective release and accumulation of high levels of miR-491-3p and -5p from BioRNA/miR-491-3p and -5p agent, respectively, not only underlines the stability and processing preference and efficiency of the hsa-pre-miR-34a carrier, which might be affected to certain degree by the tRNA scaffold, but also indicates the utilities of unparalleled BioRNA/miR-491-3p and -5p molecules for studying their functional differences. MiR-491-5p is widely recognized as a tumor suppressor that regulates cancer progression by targeting some anti-apoptotic and proliferative factors, such as the B-cell lymphoma-extra-large^51,52 and tumor protein p53.⁵² Further, miR-491-5p has been reported to regulate epoxide hydrolase 1,⁵³ an enzyme involved in the detoxification of xenobiotics, such as carbamazepine and phenytoin, as well as endogenous epoxides. By contrast, miR-491-3p is relatively less studied, but it has been identified to modulate some components of xenobiotic metabolism, including the UGT1A enzyme family,¹⁹ as well as the efflux transporter ABCB1 (P-glycoprotein).⁵⁴ Among the UGT1A isoforms, UGT1A1 plays a central role in glucuronidation-mediated detoxification.9, 10, 11 Indeed, our extensive biochemical and functional studies established the effectiveness of miR-491-3p in the regulation of UGT1A1 protein outcomes and xenobiotic-metabolizing capacities in cells.

One interesting and new observation from our immunoblotting studies is the difference in the size of UGT1A1 protein bands under unboiled and boiled conditions. Although the unboiled human liver cell lysate samples only display positive bands at higher molecular weight (∼150 kDa), a major band with the expected molecular weight (∼55 kDa) shows up when boiled samples are used. This pattern indicates that UGT1A1 proteins are not only present as monomers but also as higher-order oligomers. Indeed, UGT enzymes have been revealed to form homodimers, heterodimers, and even tetramers.^55,56 As the N-terminal domains play an important role in mediating UGT protein-protein interactions,57, 58, 59 some UGTs can form covalently cross-linked oligomers via intermolecular disulfide bonds.^60,61 As such, the ∼150 kDa bands likely represent the UGT1A1 higher-order oligomers, which, after boiling, the disulfide or noncovalent interactions are disrupted, revealing the monomeric form (∼55 kDa). Nevertheless, the presence of UGT1A1 oligomers warrants further validation by using other technologies, such as the MS/MS-based proteomics approach.

In line with previous finding on the reduction of UGT1A1 mRNA levels as well as raloxifene glucuronidation activities by 50 nM of miR-491-3p mimics in human Huh7 cells,¹⁹ our comprehensive studies with 20 nM of recombinant pre-miR-491, miR-491-5p, and miR-491-3p molecules demonstrated the effectiveness of miR-491-3p to suppress UGT1A1 protein levels for 20%–40% in Hep3B and HepG2 cells. Although this study does not directly compare the activities of bioengineered miR-491-3p and chemically synthesized mimics, the latter undoubtedly differ and vary largely among different vendors, our previous side-by-side comparisons have revealed greater or equal activities for our BioRNAs when compared with the same doses of chemically synthesized counterparts,^35,62,63 illustrating BioRNAs a unique addition to conventional tools for research. Efficacy of recombinant miR-491-3p to modulate UGT1A1 protein levels is not only manifested by the decrease of ∼150-kDa immunoblot bands under unboiled conditions but also the ∼55-kDa band under boiled conditions. Moreover, the reduction of UGT1A1 protein levels in Hep3B and HepG2 cells by BioRNA/miR-491-3p was visualized through the IFC confocal imaging studies, and the subcellular localization of UGT1A1 protein on ER was confirmed. Consequently, BioRNA/miR-491-3p treatment decreased UGT1A1-mediated xenobiotic metabolism capacity, as indicated by a lower fluorescent signal after NHPN incubation,⁴⁴ when compared with control RNA. Interestingly, the BioRNA^Gly/miR-491-3p was slightly more effective to regulate UGT1A1 than BioRNA^Leu/miR-491-3p, suggesting possible influence of the tRNA scaffold on BioRNA stability, processing, and thus apparent efficacy of payload miRNA.^33,35,36 Together, these findings illustrate the significance of miR-491-3p-controlled PTGR to govern cellular UGT1A1 protein levels that may be translated into meaningful variability in glucuronidation metabolism.

The impact of miR-491-3p on UGT1A1-catalyzed drug metabolism in Hep3B cells was precisely delineated in present study by using the index reaction, estradiol 3-glucuronidation,⁶⁴ and selective LC-MS/MS methods. By contrast, cellular estradiol 17-glucuronidation capacity mediated mainly by UGT1A3 and 2B7 was largely unaffected by miR-491-3p. This finding, highlighting the utilities of form-specific index reactions in drug metabolism, indicates potential role of miR-491-3p-controlled PTGR of UGT1A1 in glucuronidation of relevant endobiotics and xenobiotics as well as possible physiological, pharmacological, or toxicological consequences. Since UGT1A isoforms share a common 3′UTR,^3,6, 7, 8 one cannot exclude any possible effects of miR-491-3p on the protein levels of other UGT1A enzymes and overall UGT activity without more comprehensive studies. Indeed, previous studies have found the influence of miR-491-3p mimics on UGT1A3 and 1A6 mRNA levels, besides UGT1A1.¹⁹ Therefore, efforts were made in the present study to measure overall UGT activity with a fluorometric UGT Activity Assay/Ligand Screening Kit (Abcam) in addition to the 2 UGT1A1 index reactions. Although it does not reach statistical significance, total UGT activity was reduced for approximately 50% by recombinant miR-491-3p when compared with the control RNA (Supplemental Fig. 3). Considering the degrees of change in estradiol 3-glucuronidation activity (45%) and overall UGT activity (50%) by miR-491-3p in Hep3B cells, it is likely attributable to UGT1A1. Rather, different cell lines and tissues may exhibit different transfection efficiencies and have distinct genome and regulatory mechanisms, leading to variable levels of individual UGT1A isoforms that are differentially modulated by multiple or different factors, including miRNAs19, 20, 21, 22, 23, 24 that might be further complicated by possible compensatory mechanisms. Caution is advised to generalize observations under particular circumstances, such as in vitro to in vivo extrapolation.

In summary, this study successfully designed and produced 6 novel recombinant miR-491 molecules using 2 different human tRNA scaffolds and their fused hsa-pre-miR-34a carriers. Our results demonstrated the biogenesis of both 5p and 3p strands from BioRNA/pre-miR-491 in human liver cells, as well as selective release of payload miR-491-5p and -3p from BioRNA/miR-491-5p and -3p, respectively. Further studies established the effectiveness of BioRNA/miR-491-3p to modulate UGT1A1 protein expression and subsequently, cellular UGT1A1-catalyzed drug-metabolizing capacity, where the presence of native UGT1A1 oligomers was also noted in cells. Overall, these findings highlight the significance of PTGR mechanisms in drug metabolism and potential of recombinant miRNA molecules.

Conflict of interest

The authors declare no conflicts of interest.